OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Evaluation & Observability

Seqeval

Seqeval is a Python library that evaluates sequence-labeling and chunking predictions against reference labels, returning metrics and classification reports for tasks including named-entity recognition, part-of-speech tagging and semantic role labeling.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Seqeval — OpenViST project profile in AI Evaluation & Observability.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Seqeval after a model or other system has produced token-level labels, when you need metrics or reports and want to select between the default conlleval-compatible evaluation and strict evaluation for a tagging scheme.

Who it is for

  • Developers evaluating named-entity recognition or other sequence-labeling systems.
  • Researchers comparing token-label predictions with reference annotations.
  • Teams that need classification metrics under a specified tagging scheme.

Documented uses

  • Calculate accuracy, precision, recall or F1 for predicted token labels.
  • Generate a text classification report for sequence-labeling results.
  • Compare evaluation under the default conlleval-compatible mode and strict scheme-based mode.
Before you choose

Requirements & limits

  • Predictions and reference labels must be supplied as sequences of tags.
  • Default and strict modes can score the same labels differently; strict mode requires a scheme argument.
  • The repository identifies the library as MIT-licensed.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Seqeval and evaluate labeled sequences

Install the library and calculate metrics or a classification report for reference and predicted token labels.

Before you start

  • Python environment with pip
  • Reference and predicted labels arranged as sequences of tags

Steps

  1. Install the package with the command below.

    Documented command
    pip install seqeval

  2. Import a metric such as f1_score or classification_report from seqeval.metrics.

  3. Pass the reference labels and predictions to the selected metric; use the default mode or, for strict evaluation, provide mode='strict' and a scheme such as IOB2.

  4. Read the resulting score or classification report.

Expected outcome

A sequence-labeling metric or report for the supplied reference and predicted tags.

Guide limitations

  • Default mode is compatible with conlleval and can produce different results from strict evaluation.
  • Strict mode requires both mode='strict' and a scheme argument.
  • IOBES and BILOU are supported only in strict mode.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
chakki-works/seqeval · Official homepage title · retained in its source language
How we review sourcesSuggest a correction